Semiconductors are essential for most industries and are ubiquitous in the tools and hardware of the modern economy. During the COVID-19 pandemic, widespread take-up of remote working caused demand for this hardware to surge just as supply-chain disruptions interrupted the production of chips. This has led to a global chip shortage, damaging industries reliant on them.
Demand for semiconductors is expected to continue to grow. Creating a resilient supply chain for this essential good is therefore a crucial challenge. The industry is better equipped than government to match supply with demand, but geostrategic considerations justify government chip strategies. In February 2022, the European Commission proposed a European Chips Act.
Like the CHIPS and Science Act in the United States and China’s Big Fund for semiconductors, the European Chips Act would include major subsidies for semiconductor production. Smaller amounts would be allocated to chip design and other high-end tasks essential for semiconductor production. But the European Chips Act should also be designed to strengthen the EU’s position in semiconductor research and development, and to improve market conditions for new entrants and incumbents alike, facilitating a greater role for Europe in the global chip industry.
Europe’s minor role
Europe does not play a major role in the market for chips, in which the manufacturing leaders are Taiwan and South Korea. Semiconductors are mostly used for the production of digital hardware, which the EU imports as final goods. Europe produces only 10% of the world’s chips, primarily for industrial and automotive applications. So far, Europe has no capacity to produce cutting-edge chips and plays a minor role in computer chip design, accounting for only 2% of the market for chip design outside of companies that both design and produce chips.
However, the EU is significant in upstream parts of the global chip ecosystem. The most important producer of manufacturing equipment used in foundries – where chips are made – is Dutch company ASML which has a monopoly in the high-end market. Many of its suppliers are also European. Europe is also present in high-end design, with the IMEC in Belgium considered the most important research institute.
Current chip production capacity reflects the demand that was anticipated a few years ago before the demand surge from the pandemic. New foundries take years to build and can cost over $10 billion. Switching production between different chips to adapt to shifting demand can take months. Companies are already investing in new manufacturing facilities, but increased capacity will take time.
Governments are seeking to gain market share and reduce their dependence on this important supply chain, with total government support for the chips industry amounting to 0.9% of 2020 global GDP. A catalyst for this was China’s landmark industrial policy strategy, Made in China 2025. Introduced in 2015, it created two major funds with a total of 1 trillion renminbi (around $150 billion) to end China’s dependence on imported chips. The US CHIPS and Science Actincludes subsidies of $52 billion. In 2021, India announced a plan to attract chipmakers with $10 billion.
Pushing for a greater role
The European Chips Act promises €43 billion in “policy-driven investment” up to 2030. It includes some new and some repackaged funding for research and development, as well as new rules that allow EU countries to use subsidies to lure manufacturers:
- The first pillar of the Chips Act focuses funding for R&D and innovation. The EU will provide €3.3 billion of R&D and innovation funding, complemented by a €2 billion ‘Chips Fund’ for start-ups consisting of leveraged European Investment Bank and EU finance. Together with money from EU countries for industrial innovation (part of a so-called Important Project of Common European Interest), this amounts to an overall budget for the R&D and innovation ‘pillar’ of the Chips Act of €11 billion.
- The second pillar introduces a new state-aid exemption that will allow EU countries to finance the construction of foundries that use technologies new to the EU. Industrial subsidies under this pillar will account for most of the €43bn headline number and will be financed by national governments. The focus will be on attracting the three cutting-edge chip manufacturers (TSMC, Samsung and Intel). This is an expensive endeavour. Reported German government subsidies for a new Intel foundry in Magdeburg alone are €6.8 billion, twice the amount budgeted by the EU for R&D under the Chips Act’s first pillar.
- The R&D support and industrial subsidies meant to increase capacity will be complemented by new emergency measures under the third pillar. These are meant to ensure supply in times of shortage and include joint procurement of chips by the European Commission, ‘EU first’ provisions for foundries that benefited from subsidies and export controls.
All in all, and even if the investment is smaller than China or the US (when taking into account subsidies by states), the European Chips Act marks a turning point for European industrial policy. The EU has strong state-aid rules that are intended to prevent subsidy races within the EU. The Chips Act embraces industrial policy in a key sector. While meeting Europe’s chips needs is important in a new geopolitical contect, design of an EU-level industrial policy for chips should focus on the highest end of the semiconductor value chain, and especially on R&D instead of production subsidies. Focusing on industrial subsidies so that national governments can use new state-aid exemptions to reach the Chips Act goals is not necessarily the best path for the EU to become a more significant player in the semiconductor value chain. The strategy risks pitting EU countries against each other in a competition over subsidies. Since funding for these subsidies comes from national budgets, it is unsurprising that it was the richest member (Germany) that announced the first large planned investment linked to the Chips Act – the Magdeburg foundry – while poorer EU countries risk being left behind.
Meeting EU needs?
Will the substantial public funding promised by the Chips Act and the shift in industrial policy it signifies serve European needs? Are the goals of increasing the resilience of European supply against shortages and geopolitical intervention the right ones?
On protecting against shortages, the proposed Chips Act seems ill-suited. While the EU imports goods that have chips in them, it does not import large volumes of chips themselves. The shortages that plague the automotive industry do not affect the cutting-edge segment of the market that will receive most of the Chips Act funding. Therefore, as import substitution or reshoring policy, the Chips Act will not target current EU demand, and the new capacity it promotes will also not address the current shortages because foundries take years to build. Meanwhile, it is not clear if government interventions will be able to prevent future shortages, or if intervention will be efficient if more shortages occur. There is no reason why governments should be better than markets at managing supply and demand. The incentives for the private sector to build sufficient capacity are clear, but given the high costs and long construction time for foundries mismatches between supply and demand are bound to happen.
The efficacy of the ‘emergency’ measures proposed in the Chips Act (the third pillar) is also questionable. The measures assume wrongly that the government could redistribute highly specialised chips between different end-users. Export controls and disruptions of supply chains through domestic prioritisation, as foreseen by the proposed Chips Act, have the potential to further distort markets in times of shortages.
However, despite all these shortcomings, a comprehensive EU strategy on semiconductors is needed because of the industry’s economic and geopolitical importance. The EU’s energy transition and digitalisation plans are highly dependent on the availability of chips. The reliance on just a few geographical areas – the US for design, Taiwan and South Korea for fabrication, especially of advanced chips – is problematic, as it exposes one of the most important global value chains to local shocks (both environmental and political). Finally, growing risks of global market fragmentation and political instability make it even more important to react now as an important step towards Europe’s strategic autonomy.
Control of semiconductor technologies is also an important deterrent against economic coercion by third parties. The solution is to design an effective industrial policy for this sector that serves the EU’s needs in a cost-effective way. This means spending in the highest end of this sector and mostly in R&D rather than in production of easily reproducable chips where overcapacity is looming.
The Chips Act’s focus on subsidies for high-end manufacturing is misguided because the EU does not necessarily need all types of advanced semiconductors but only some, based on its industrial comparative advantage. Better coordination of support programmes with partner countries outside the EU would allow the public costs of diversifying the production of chips to be reduced. There are talks on avoiding subsidy races, but these are not stopping subsidy programmes from moving ahead. Government programmes for foundries in addition to massive private investment create overcapacity risks. Competing subsidy programmes could lead to trade conflicts between the EU and the US at the World Trade Organisation. Large domestic capacity is unnecessary when in control of key technologies. Finally, there is also no need to produce advanced chips because to control part the value chain is enough, as shown by the US. The US has imposed export controls on chips and chip components exported to China and Russia as part of sanctions over Russia’s invasion of Ukraine. In doing so, it has relied on its leading role in chip design, and not on its relatively small domestic production (which is of a scale comparable to the EU).
All in all, there is no doubt than in a more geopolitical world, industrial policy considerations become more important. The EU needs industrial policy in key sectors where supply bottlenecks are a real possibility. The chip industry is probably the most relevant. However, the objective of such industrial policy, as embodied in the proposed European Chips Act, should be on developing those parts of the semiconductor industry where the EU can gain strategic leverage, while continuing to improve overall market conditions. More specifically, while Europe hosts leading chip research, it fails to commercialise European innovations. The R&D and innovation part of the Chips Act includes many initiatives to address this, including financial support for start-ups. However, the funding is tilted towards industrial subsidies to build capacity. Fabrication is the phase in the chips value chain that requires most fixed asset investment and, thus, more subsidies but it is not the only relevant one. Heavy subsidies to produce chips that EU industries do not use will be expensive and less useful then increasing Europe’s role in design by investing in R&D. Finally, improving market conditions could also be done without jeopardising the level playing field and undermining the EU’s credibility when pushing for a level playing field internationally.